7 Essential Linux Tweaks to Supercharge Redis Performance

Redis developers often focus on Redis‑specific configuration, but operating‑system settings are crucial for stable, high‑performance operation.

1. Memory allocation control

vm.overcommit_memory

When vm.overcommit_memory is set to 0, the kernel may reject memory allocation, causing fork failures during background saves. Setting it to 1 allows overcommit and prevents these errors.

# WARNING overcommit_memory is set to 0! Background save may fail under low memory condition. To fix this issue add 'vm.overcommit_memory = 1' to /etc/sysctl.conf and then reboot or run the command sysctl vm.overcommit_memory=1 for this to take effect.

Check the current value: cat /proc/sys/vm/overcommit_memory Set the value:

echo "vm.overcommit_memory=1" >> /etc/sysctl.conf
sysctl vm.overcommit_memory=1

Configure a reasonable maxmemory and keep 20‑30% of RAM free.

Centralize AOF rewrite and RDB bgsave management.

Set vm.overcommit_memory=1 to avoid fork failures.

2. Swappiness

Parameter meaning

Swappiness (0‑100) controls how aggressively Linux uses swap. Lower values keep data in RAM; higher values increase swap usage. The default is 60. For Redis, values 0‑1 are recommended on kernels >3.5, otherwise 0.

Configuration

echo {bestvalue} > /proc/sys/vm/swappiness

Persist across reboots:

echo vm.swappiness={bestvalue} >> /etc/sysctl.conf

Monitoring swap

Use free -m to view total swap usage, vmstat 1 for real‑time swap activity, and inspect /proc/<pid>/smaps for per‑process swap consumption.

3. Transparent Huge Pages (THP)

THP can increase latency and memory usage for Redis. Disable it:

echo never > /sys/kernel/mm/transparent_hugepage/enabled

Persist the setting by adding the command to /etc/rc.local. On some distributions the path is /sys/kernel/mm/redhat_transparent_hugepage/enabled.

4. OOM killer

Adjust oom_adj for Redis processes to a low value (e.g., -17) to reduce the chance of being killed when memory is scarce.

for redis_pid in $(pgrep -f "redis-server"); do
  echo -17 > /proc/${redis_pid}/oom_adj
done

5. NTP

Synchronize system clocks to avoid time drift between Redis nodes, which simplifies troubleshooting of Sentinel or Cluster events.

0 * * * * /usr/sbin/ntpdate ntp.xx.com > /dev/null 2>&1

6. ulimit

Increase the open‑files limit to accommodate maxclients. For example, set ulimit -Sn 10032 or edit /etc/security/limits.conf to raise the soft limit.

7. TCP backlog

Ensure /proc/sys/net/core/somaxconn is at least the Redis tcp-backlog (default 511). If it is lower, Redis will emit a warning.

echo 511 > /proc/sys/net/core/somaxconn